Prepare 0.2.0 CRAN resubmission

[d-morrison]

Summary

rwicc was archived from CRAN on 2026-01-30 because it required the archived package pryr (already replaced with lobstr in #11). This PR prepares the 0.2.0 release to restore it, following https://r-pkgs.org/release.html.

• Bump version 0.1.3.9006 → 0.2.0.
• Finalize NEWS.md under the 0.2.0 heading (lobstr dependency change, graph_omega() + graph_simulated_*() helpers, the graph_S/fit_joint_model/simulate_interval_censoring fixes); dropped dev-tooling entries.
• Bind the two NSE column names so R CMD check --as-cran drops the "no visible binding" note — now 0 errors / 0 warnings / 1 note (the expected CRAN-incoming/archived note).
• Use the canonical doi.org form for the paper link in the README.
• Rewrite cran-comments.md for the resubmission (archival cause/fix + the two expected URL notes).

Local check results

• rcmdcheck::rcmdcheck(args = c("--no-manual", "--as-cran")): 0 errors | 0 warnings | 1 note (CRAN-incoming: new submission / previously archived; two README URLs flagged — the Wiley DOI 403 and the CRAN-checks badge 404, both expected and explained in cran-comments).
• urlchecker::url_check(): only the two expected URLs above.
• spelling::spell_check_package(): clean.

Remaining (maintainer) steps before submitting to CRAN

• win-builder (devel + release) and R-hub cross-platform checks (being triggered).
• devtools::submit_cran() once those return clean.

🤖 Generated with Claude Code

[claude]

Claude finished @d-morrison's task in 2m 6s —— View job

---

Code Review: PR #32 — Prepare 0.2.0 CRAN resubmission

• Gather context (changed files, diff)
• Read all 7 changed files
• Analyze R code changes
• Post review feedback

---

This is a clean, well-scoped release-prep PR. The changes are minimal and purposeful. A few observations:

---

R/fit_joint_model.R — NSE binding

The binding at line 125 correctly suppresses the R CMD check note:

`P(S=s|e,l,r,o,y)` <- NULL

Minor comment inaccuracy: the comment says "referenced as a bare symbol in the bigglm/glm weight = formula below", but the bigglm call at line 495 uses weight = ~\P(S=s|e,l,r,o,y)`(a tilde formula, not a bare symbol). Only thestats::glm()fallback at line 506 uses it as a bare symbol and triggers the note. The comment could simply sayglminstead ofbigglm/glm` to be precise, but this is trivial. Fix this →

---

R/simulate_interval_censoring.R — NSE binding

The new `years from study start to seroconversion` <- NULL at line 85 is correct — the column is used as a bare symbol at lines 119, 127, 201, and 218. Properly added to the existing binding block. ✓

---

README.Rmd + README.md — DOI canonicalization

Both files updated consistently: onlinelibrary.wiley.com → doi.org canonical DOI. The README.md diff (+1/-2) is consistent with regeneration from README.Rmd (the old content was split across two lines; the new content fits on one). ✓

---

NEWS.md

Heading finalized from (development version) to 0.2.0. Dev-tooling entry (reprexes skill docs) correctly removed — it was never user-facing. Remaining entries are accurate and appropriately detailed. ✓

---

DESCRIPTION

Version bump 0.1.3.9006 → 0.2.0 is correct. No new dependencies. ✓

---

cran-comments.md

Well-structured resubmission note. Archival cause and fix are clearly stated. The two expected URL notes (Wiley 403, CRAN-checks 404) are pre-explained. Test environments cover win-builder, R-hub, and GHA matrix. ✓

---

Checklist against custom review criteria

Criterion	Status
Statistical/numerical logic unchanged	✓ No behavioral changes
No partial argument matching introduced	✓
New exported functions have full roxygen docs	N/A — no new exports
Conforms to .lintr settings	✓ (<- NULL NSE pattern is idiomatic)
No unjustified new imports in DESCRIPTION	✓ lobstr already present from #11
Behavior changes come with tests	N/A — NSE bindings are check-note suppression only
README.md regenerated from README.Rmd	✓
Version bumped + NEWS entry	✓
CI workflows unchanged	✓

Overall: looks ready for devtools::submit_cran() once win-builder and R-hub results are clean. The one minor comment inaccuracy in fit_joint_model.R is not a blocker.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Files with missing lines	Coverage Δ
R/fit_joint_model.R	81.41% <100.00%> (+0.06%)	⬆️
R/simulate_interval_censoring.R	100.00% <100.00%> (ø)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[claude]

Claude finished @d-morrison's task in 1m 51s —— View job

---

Code Review: PR #32 — Prepare 0.2.0 CRAN resubmission

• Gather context (CLAUDE.md, diff, changed files)
• Read all 7 changed files
• Analyze R code changes
• Post review feedback

---

This is a clean, minimal release-prep PR. Changes are purposeful and well-scoped. One minor comment inaccuracy, no blockers.

---

R/fit_joint_model.R — NSE binding

The binding at line 125 is correct and necessary. One minor inaccuracy in the comment:

# bind the non-standard column name referenced as a bare symbol in the
# bigglm/glm `weight =` formula below, to avoid an R CMD check
# "no visible binding for global variable" note:
`P(S=s|e,l,r,o,y)` <- NULL

The biglm::bigglm() call at line 495 uses weight = ~\P(S=s|e,l,r,o,y)`— a **tilde formula**, not a bare symbol. Only thestats::glm()fallback at line 506 usesweight = `P(S=s|e,l,r,o,y)`as a bare symbol and triggers the note. The comment should saystats::glmrather thanbigglm/glm`. Trivial, not a blocker. Fix this →

---

R/simulate_interval_censoring.R — NSE binding

`years from study start to seroconversion` <- NULL at line 85 is correct. The symbol is used as a bare symbol at line 218 (elapsed time = ... - `years from study start to seroconversion`) and in .by string at line 201. Properly appended to the existing binding block. ✓

---

DESCRIPTION

• Version bump 0.1.3.9006 → 0.2.0 ✓
• R (>= 4.1.0): correct — the base pipe |> (R 4.1+) is used throughout the package ✓
• LazyData: true removed: correct — there is no data/ directory; this was a spurious field that would generate an R CMD check NOTE ✓
• No new imports ✓

---

NEWS.md

Heading finalized from (development version) to 0.2.0. New entry for the R (>= 4.1.0) requirement added. The dev-tooling entry for the reprexes skill docs was correctly dropped (never user-facing). Remaining entries are accurate. ✓

---

README.Rmd + README.md

Both files updated consistently: onlinelibrary.wiley.com/doi/... → canonical doi.org/10.1111/biom.13472 form. The README.md diff is consistent with knitr regeneration from README.Rmd. ✓

---

cran-comments.md

Well-structured resubmission note. Archival cause (pryr archived) and fix (lobstr) are clearly stated. The two expected URL notes (Wiley 403, CRAN-checks badge 404) are pre-explained with context. Test environments are comprehensive. ✓

---

Checklist against custom review criteria

Criterion	Status
Statistical / numerical logic unchanged	✓ No behavioral changes
No partial argument matching introduced	✓
New exported functions have full roxygen docs	N/A — no new exports
Conforms to lintr style (snake_case, line length, object usage)	✓ (<- NULL NSE pattern is idiomatic)
No unjustified new imports in DESCRIPTION	✓ lobstr already present from #11
Behavior changes come with testthat tests	N/A — NSE bindings are check-note suppression only
README.md regenerated from README.Rmd	✓
Version bumped + NEWS entry	✓
LazyData field consistent with package structure	✓ (spurious field removed)

Overall: looks ready for devtools::submit_cran() once win-builder and R-hub return clean. The one minor comment inaccuracy in fit_joint_model.R is not a blocker.

[claude]

Claude finished @d-morrison's task in 2m 1s —— View job

---

Code Review: PR #32 — Prepare 0.2.0 CRAN resubmission

• Gather context (diff, changed files)
• Read all 7 changed files
• Analyze R code changes
• Post review feedback

---

This is a clean, minimal release-prep PR. The most recent commit adds the "et al." false-positive note to cran-comments.md — that's the only change since the previous two review passes.

---

cran-comments.md — "et al." spelling note (new in this sync)

The note is correct: inst/WORDLIST already contains both et and al, so spelling::spell_check_package() is clean locally. The cran-comments note accurately explains this as a CRAN incoming-feasibility checker quirk (not a local R CMD check issue). ✓

---

R/fit_joint_model.R — NSE binding comment (carried over from prior reviews)

Still a minor inaccuracy in the comment at lines 122–124:

# bind the non-standard column name referenced as a bare symbol in the
# bigglm/glm `weight =` formula below, to avoid an R CMD check
# "no visible binding for global variable" note:

biglm::bigglm() at line 495 uses weight = ~\P(S=s|e,l,r,o,y)`— a tilde formula, not a bare symbol. Onlystats::glm()at line 506 usesweight = `P(S=s|e,l,r,o,y)`as a bare symbol and triggers the note. The comment should saystats::glmrather thanbigglm/glm`. Not a blocker. Fix this →

---

All other changes — confirmed clean

File	Change	Status
DESCRIPTION	0.1.3.9006 → 0.2.0; R >= 4.0.0 → 4.1.0; LazyData removed	✓ — base pipe |> used 96× confirms 4.1.0; no data/ directory confirms LazyData removal
NEWS.md	Heading finalized; R >= 4.1.0 entry added; dev-tooling entry dropped	✓
R/simulate_interval_censoring.R	`years from study start to seroconversion` <- NULL	✓ — bare symbol at line 218 confirms binding is necessary
README.Rmd + README.md	onlinelibrary.wiley.com → doi.org canonical form; consistent regeneration	✓
cran-comments.md	Archival cause/fix, et al. note, two URL notes, test environments, reverse deps	✓

Overall: ready for devtools::submit_cran() once win-builder and R-hub return clean. The one minor comment inaccuracy in fit_joint_model.R is not a blocker.

[claude]

Claude finished @d-morrison's task in 1m 58s —— View job

---

Code Review: PR #32 — Prepare 0.2.0 CRAN resubmission

• Gather context (diff, changed files)
• Read changed files
• Analyze changes since last review pass
• Post review feedback

---

This sync adds one new commit (84cf2d0 add positai files) on top of the previously reviewed state. All prior findings remain unchanged; one new item to note.

---

New in this sync: .positai ignore entries

.Rbuildignore and .gitignore both get a ^\.positai$ / .positai entry. Both are correct:

• .gitignore keeps the Posit AI tool config out of version control. ✓
• .Rbuildignore keeps it out of the package tarball (necessary even for gitignored files that may be present locally during R CMD build). ✓

---

Carried from prior reviews: R/fit_joint_model.R comment inaccuracy (non-blocker)

The comment at lines 122–124 says "bigglm/glm weight = formula below", but:

• biglm::bigglm() at line 495 uses weight = ~\P(S=s|e,l,r,o,y)`` — a tilde formula, not a bare symbol; R CMD CHECK does not flag formula contexts.
• stats::glm() at line 506 uses weight = \P(S=s|e,l,r,o,y)`` as a bare symbol — this is what triggers the note.

The comment should say stats::glm rather than bigglm/glm. Trivial, not a blocker. Fix this →

---

All other changes — confirmed clean

File	Change	Status
DESCRIPTION	0.1.3.9006 → 0.2.0; R >= 4.0.0 → 4.1.0; LazyData removed	✓
NEWS.md	Heading finalized; R >= 4.1.0 entry added; dev-tooling entry dropped	✓
R/fit_joint_model.R	`` `P(S=s	e,l,r,o,y)` <- NULL `` NSE binding
R/simulate_interval_censoring.R	`years from study start to seroconversion` <- NULL	✓ — bare symbol at line 218
README.Rmd + README.md	onlinelibrary.wiley.com → doi.org canonical DOI; consistent knitr regen	✓
cran-comments.md	Archival cause/fix, "et al." note, two URL notes, test envs, reverse deps	✓

Overall: clean and ready for devtools::submit_cran() once win-builder and R-hub return clean. The comment inaccuracy in fit_joint_model.R is the only open finding and is not a blocker.

[claude]

Claude finished @d-morrison's task in 1m 56s —— View job

---

Code Review: PR #32 — Prepare 0.2.0 CRAN resubmission

• Gather context (diff, changed files)
• Read CLAUDE.md and changed files
• Analyze new commits since last review
• Post review feedback

---

The latest commit (087f740) bumps the version to 0.2.0.9000. One new blocker to flag; all other prior findings unchanged.

---

DESCRIPTION — version bump to 0.2.0.9000 (new — blocker)

Version: 0.2.0.9000 is a development version (the usethis post-release convention). CRAN requires a release version — no fourth component or .9000 suffix. Running devtools::submit_cran() against this DESCRIPTION will warn or fail:

Version: 0.2.0.9000   ← wrong for a CRAN submission

The standard workflow is:

1. Submit 0.2.0 to CRAN.
2. After CRAN acceptance, bump to 0.2.0.9000.

If the submission hasn't happened yet, revert to 0.2.0 before running devtools::submit_cran(). If it already happened and this bump is the post-acceptance follow-up, no action needed — but in that case the PR title/description should say so. Fix this →

---

NEWS.md — empty (development version) section (related)

After the version bump, NEWS.md now begins:

# rwicc (development version)

# rwicc 0.2.0
[entries]

This empty leading section is the standard usethis::use_dev_version() output — correct if this is the post-release bump, but jarring on a CRAN-submission PR. If reverting to 0.2.0, also remove the empty (development version) heading.

---

R/fit_joint_model.R — comment inaccuracy (carried from prior reviews, non-blocker)

# bind the non-standard column name referenced as a bare symbol in the
# bigglm/glm `weight =` formula below, to avoid an R CMD check

biglm::bigglm() (line 495) uses weight = ~\P(S=s|e,l,r,o,y)`— a tilde formula, not a bare symbol. Onlystats::glm()(line 506) uses it as a bare symbol and triggers the note. Should saystats::glmrather thanbigglm/glm`. Trivial. Fix this →

---

All other changes — confirmed clean

File	Change	Status
.Rbuildignore / .gitignore	^\.positai$ / .positai ignore entries	✓
R/simulate_interval_censoring.R	`years from study start to seroconversion` <- NULL	✓
README.Rmd + README.md	onlinelibrary.wiley.com → doi.org canonical DOI; consistent regen	✓
cran-comments.md	Archival cause/fix, "et al." note, two URL notes, test envs, reverse deps	✓

---

Overall: one blocker — Version: 0.2.0.9000 should not be what goes to CRAN. If the submission already happened externally, this note is moot; if not, revert to 0.2.0 (and remove the empty NEWS section) before devtools::submit_cran().